| *** asettle has joined #openstack-swift | 00:02 | |
| notmyname | backports approved. the existing +1s from other cores there was very helpful | 00:06 |
|---|---|---|
| notmyname | (I hope that the order I went through them doesn't cause merge issues) | 00:06 |
| *** dmorita has quit IRC | 00:06 | |
| *** asettle has quit IRC | 00:07 | |
| *** dmorita has joined #openstack-swift | 00:07 | |
| clayg | notmyname: hello! | 00:15 |
| clayg | notmyname: did i read your tweet right? SLC (!?) | 00:15 |
| notmyname | salt lake city is cold! | 00:15 |
| notmyname | it's currently 23! Fahrenheit! | 00:15 |
| notmyname | low tonight is 10 | 00:15 |
| notmyname | ten | 00:15 |
| clayg | scoff - global "warming" | 00:15 |
| notmyname | I saw an article that claims it's colder now that it was in june! (although mattoliverau wouldn't agree with that statement) | 00:16 |
| mattoliverau | wow, yeah, not that cold here. esp when compared to June. Seeing as june is winter :P | 00:18 |
| notmyname | :-) | 00:19 |
| mattoliverau | sigh, ssbench was killed over night due to OOM killer. Time to make a cloud server with a little more ram to be the master. Also, I had to hack ssbench to allow me to send container headers to put (turn on sharding).. so will add some tests and push that up at some stage. | 00:20 |
| notmyname | oh, cool | 00:21 |
| *** dmorita has quit IRC | 00:21 | |
| notmyname | yeah, that would be good | 00:21 |
| notmyname | were you able to use the same mechanisms used for sending the policy header? | 00:21 |
| *** dmorita has joined #openstack-swift | 00:22 | |
| mattoliverau | in a way, the policy is sent in as a param. But now you can specify container_put_headers in your scenario files. | 00:22 |
| mattoliverau | I might clean it up a little. its more just to get ssbench to do my bidding :P | 00:23 |
| *** catintheroof has quit IRC | 00:23 | |
| *** catintheroof has joined #openstack-swift | 00:24 | |
| *** dmorita has quit IRC | 00:26 | |
| *** dmorita has joined #openstack-swift | 00:28 | |
| *** catintheroof has quit IRC | 00:28 | |
| *** _JZ_ has quit IRC | 00:32 | |
| *** manous has joined #openstack-swift | 00:46 | |
| clayg | kota_: patch 393263 is still looking good!!! | 00:46 |
| patchbot | https://review.openstack.org/#/c/393263/ - liberasurecode - ISA-L Cauchy support | 00:46 |
| clayg | cschwede: friendly ping to add patch 346475 to loop back to patch on your list cc: torgomatic timburke anyone else interested in rings | 00:49 |
| patchbot | https://review.openstack.org/#/c/346475/ - swift - For any part, only one replica can move in a rebal... | 00:49 |
| *** tovin07 has joined #openstack-swift | 00:52 | |
| clayg | yay tsg and kevin are both making comments on the ISA-L change (!?) | 00:57 |
| *** arch-nemesis has quit IRC | 00:58 | |
| kota_ | good morning | 01:01 |
| mattoliverau | kota_: morning | 01:01 |
| kota_ | clayg: thanks for reviewing and tsg and kmgreen2 make small comments, i will update soon to land quickly | 01:02 |
| *** asettle has joined #openstack-swift | 01:03 | |
| *** Jeffrey4l has quit IRC | 01:06 | |
| *** Jeffrey4l has joined #openstack-swift | 01:06 | |
| *** jamielennox is now known as jamielennox|away | 01:07 | |
| *** asettle has quit IRC | 01:07 | |
| *** Jeffrey4l has quit IRC | 01:08 | |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/liberasurecode: ISA-L Cauchy support https://review.openstack.org/393263 | 01:09 |
| *** Jeffrey4l has joined #openstack-swift | 01:09 | |
| kota_ | mattoliverau: morning | 01:09 |
| kota_ | clayg: done | 01:09 |
| kota_ | hope that can land soon | 01:09 |
| *** vinsh has joined #openstack-swift | 01:19 | |
| *** Jeffrey4l has quit IRC | 01:20 | |
| *** Jeffrey4l has joined #openstack-swift | 01:21 | |
| *** jamielennox|away is now known as jamielennox | 01:21 | |
| *** david-lyle has joined #openstack-swift | 01:33 | |
| kota_ | clayg, timburke: it look like your turn to land patch 393263 ;-) | 01:37 |
| patchbot | https://review.openstack.org/#/c/393263/ - liberasurecode - ISA-L Cauchy support | 01:37 |
| clayg | kota_: it's too bad maybe that isa-l is not tested in the gate? not sure... | 01:37 |
| *** Jeffrey4l has quit IRC | 01:39 | |
| kota_ | clayg: exactly, we cannot test for now | 01:39 |
| *** Jeffrey4l has joined #openstack-swift | 01:39 | |
| kota_ | clayg: i think if we will have sort of test_setup.sh supported by infra team, I'd like to make a script to setup isa-l from source | 01:39 |
| kota_ | clayg: or if you know how to run some scripts with sudo before testing in the gate, please let me know, that's a way I'm looking for. | 01:41 |
| kota_ | Or could we have custom test script for c lang project? | 01:42 |
| *** mingyu has joined #openstack-swift | 01:43 | |
| kota_ | it might be a time to look at project-config - Or asking tdasilva | 01:43 |
| kota_ | clayg: note that, at least, I ran test ISA-L Cauchy patch with the environ isal installed and all test passed as green. So I'm sure it's nice. | 01:44 |
| timburke | i guess we go edit https://github.com/openstack-infra/project-config/blob/master/jenkins/jobs/liberasurecode.yaml ? | 01:45 |
| kota_ | timburke: nice, it looks like we can add isa-l installation there. | 01:47 |
| *** mingyu has quit IRC | 01:49 | |
| *** mingyu has joined #openstack-swift | 01:51 | |
| *** StraubTW has joined #openstack-swift | 01:52 | |
| openstackgerrit | Thiago da Silva proposed openstack/swift: adding unit tests for if-none-match fix https://review.openstack.org/401044 | 01:52 |
| kota_ | clayg: to consider you want to use ISA-L Cauchy asap, it may be better that the setting for isa-l testing land after ISA-L Cauchy land. Otherwise, probably dependency chain will stand up against to us to block ISA-L likey 'older version doesn't support the ISA-L cauchy and we cannot test that so -1 (from jenkins)' | 01:54 |
| kota_ | complicated dependency chain between c - python problem :/ | 01:55 |
| kota_ | ah, could we add building script liberasurecode from source as well as pyeclib? | 01:56 |
| *** StraubTW has quit IRC | 01:56 | |
| kota_ | um... probably it's no because we set *python-job* for pyeclib so no hooking point to install liberasurecode from source different from liberasurecode :/ | 01:57 |
| clayg | kota_: I think i said to john that I'd like one pyeclib job that tests from packaged distro liberasure.so and other that tests on built from source (maybe?) | 01:58 |
| clayg | kota_: another option/idea is on one of those also go ahead and check out master isa-l (maybe?) | 01:58 |
| clayg | kota_: none of this is a blocker for landing isa-l-rs-cauchy (because I do not want to wait!) | 01:58 |
| timburke | the split pyeclib jobs might be doable via tox envs. is there a way for us to tell the pyeclib compile about some user-provided libs so we don't have to go install liberasurecode before dropping sudo? | 02:00 |
| kota_ | tibmurke: I don't hink we can go install c library via tox w/o sudo | 02:02 |
| kota_ | timburke: IIRC, it's the resaon why we split the repo pyeclib/liberasurecode | 02:02 |
| *** asettle has joined #openstack-swift | 02:04 | |
| kota_ | timburke: say cannot but it may not impossible but it should be too tricky to install c lib via tox | 02:04 |
| kota_ | timburke: Or great idea for the way is always welcome | 02:05 |
| *** Jeffrey4l has quit IRC | 02:06 | |
| *** Jeffrey4l has joined #openstack-swift | 02:07 | |
| *** asettle has quit IRC | 02:08 | |
| *** sanchitmalhotra has quit IRC | 02:12 | |
| clayg | kota_: ok, down to patch 393276 now! | 02:28 |
| patchbot | https://review.openstack.org/#/c/393276/ - pyeclib - ISA-L Cauchy support | 02:28 |
| kota_ | clayg: nice! | 02:29 |
| *** StraubTW has joined #openstack-swift | 02:32 | |
| openstackgerrit | Merged openstack/liberasurecode: ISA-L Cauchy support https://review.openstack.org/393263 | 02:34 |
| kota_ | yey! | 02:34 |
| *** hogepodge has quit IRC | 02:38 | |
| *** m_kazuhiro has joined #openstack-swift | 02:45 | |
| clayg | for someone running liberasurecode *not* built from source (using distro packages) | 02:47 |
| clayg | it might be interesting to try out patch 393276 | 02:47 |
| patchbot | https://review.openstack.org/#/c/393276/ - pyeclib - ISA-L Cauchy support | 02:47 |
| clayg | it's just a little change to pyeclib - but you could totally clone it and pip install it and try running your swifts - everything should work - but it might be a good introduction to "all this crazy pyeclib/liberasurecode/isa-l stuff" if you were looking to dip a toe in! | 02:48 |
| *** asettle has joined #openstack-swift | 03:04 | |
| *** mingyu has quit IRC | 03:07 | |
| *** asettle has quit IRC | 03:09 | |
| *** links has joined #openstack-swift | 03:22 | |
| *** Jeffrey4l has quit IRC | 03:23 | |
| *** kei_yama has quit IRC | 03:23 | |
| *** Jeffrey4l has joined #openstack-swift | 03:24 | |
| openstackgerrit | Tim Burke proposed openstack/pyeclib: Add tox environment to test against liberasurecode master https://review.openstack.org/407829 | 03:25 |
| kota_ | timburke: oh, is that a way? | 03:28 |
| *** winggundamth_ has joined #openstack-swift | 03:31 | |
| clayg | kota_: he's explaining it to me | 03:32 |
| openstackgerrit | Kota Tsuyuzaki proposed openstack/pyeclib: Make hard-coded dependency for liberasurecode>=1.3.1 https://review.openstack.org/395998 | 03:32 |
| kota_ | trying to rebase my had-corded dependency patch to that. | 03:32 |
| clayg | it seems to let you set a env var and that plus a script will do a user space install that "works" because of crazy shit in setup.py that tells the extension about some library and include paths? | 03:32 |
| clayg | i don't understand how c works | 03:32 |
| kota_ | maybe? | 03:35 |
| *** m_kazuhiro has quit IRC | 03:36 | |
| kota_ | not sure for now, the lib looks like to be installed to user space but how we could notify the lib path for dynamic linking | 03:36 |
| *** dmorita has quit IRC | 03:37 | |
| jrichli | notmyname: I will be in Chicago next week where it is expected that lows remain in the teens. Same week in Austin, lows don't go below 45 next week. | 03:37 |
| clayg | kota_: basically timburke says you can bring up a fresh vm with all the stuff in bindep (but *not* liberasurecode) and when you tox -e liberasurecode-master it works | 03:37 |
| kota_ | clayg: hmmm | 03:38 |
| *** Jeffrey4l has quit IRC | 03:38 | |
| clayg | in theory you'd get a .tox/liberasurecode-master than when you import pyeclib you can sorta get the gist of it using this LIBERASURECODE_DIR path | 03:38 |
| clayg | i starred it | 03:38 |
| *** Jeffrey4l has joined #openstack-swift | 03:38 | |
| clayg | there's a bunch of moving parts I don't really know about - timburke is like a sponge tho - he knows everything | 03:39 |
| kota_ | or run sort of `ldconfig $LIBERASURECODE_DIR` to find the shared object lib? | 03:39 |
| kota_ | ok, we need to avoid liberasurecode from bindep to use liberasurecode from source | 03:42 |
| *** dmorita has joined #openstack-swift | 03:42 | |
| *** dmorita has quit IRC | 03:43 | |
| *** dmorita has joined #openstack-swift | 03:44 | |
| clayg | kota_: open question for me is how the contortions in the setup.py and virtualenv handle preferences when different versions are installed in different places | 03:47 |
| clayg | I don't even know currently how it handles the case where I have system packages installed and then also installed from source with sudo! | 03:48 |
| *** dmorita has quit IRC | 03:48 | |
| clayg | be nice to bounce the whole idea of someone in infra that may have some experience doing this elsewhere ... or not | 03:48 |
| clayg | I also don't know what tests we should run and how they should vote even if tim's change works - but it'd be good to know that it works sorta seperate from that | 03:49 |
| kota_ | AFAIK, that depends on search path order | 03:50 |
| kota_ | clayg: so if the user space is prior to system path like /usr/lib (or /usr/include for header), it could be ok to setup different version into different places. | 03:51 |
| kota_ | but order depended work could make another ploblem... | 03:52 |
| kota_ | idk, how order, vurtialenv refers the c libs but i think virtualenv doesn't take care of the system-wide setting in outside of python | 03:54 |
| kota_ | FWIW, looking at the result of http://logs.openstack.org/98/395998/5/check/gate-pyeclib-python27-ubuntu-xenial/de7474f/console.html#_2016-12-07_03_35_13_982126 | 03:55 |
| kota_ | it looks like pyeclib still refers /usr/local/lib/liberasurecode.so (installed by package repo) in current status. | 03:56 |
| kota_ | probably, we need something, (dropping liberasurecode from bindep? or ldconf to the source installed dir? not sure w/o trying) | 03:57 |
| kota_ | to enable pyeclib to refer user installed liberasurecode. | 03:57 |
| *** StraubTW has quit IRC | 03:57 | |
| *** tqtran has quit IRC | 03:58 | |
| *** asettle has joined #openstack-swift | 04:05 | |
| *** mingyu has joined #openstack-swift | 04:08 | |
| kota_ | oops, I was missing the tox environ is set as a different environ :/ | 04:09 |
| kota_ | so the hard-coded dependency should be failed | 04:09 |
| kota_ | let me check testenv:liberasurecode-master | 04:09 |
| *** asettle has quit IRC | 04:10 | |
| *** Jeffrey4l has quit IRC | 04:11 | |
| *** Jeffrey4l has joined #openstack-swift | 04:11 | |
| *** psachin has joined #openstack-swift | 04:12 | |
| *** mingyu has quit IRC | 04:12 | |
| *** links has quit IRC | 04:20 | |
| *** dmorita has joined #openstack-swift | 04:25 | |
| *** Jeffrey4l has quit IRC | 04:29 | |
| *** Jeffrey4l has joined #openstack-swift | 04:30 | |
| *** dmorita has quit IRC | 04:30 | |
| *** Jeffrey4l has quit IRC | 04:30 | |
| *** Jeffrey4l has joined #openstack-swift | 04:31 | |
| *** psachin has quit IRC | 04:32 | |
| kota_ | timburke: nice, all tests for patch 395998 passed with liberasurecode-master env so that way can enable to use the master. | 04:32 |
| patchbot | https://review.openstack.org/#/c/395998/ - pyeclib - Make hard-coded dependency for liberasurecode>=1.3.1 | 04:32 |
| kota_ | er, it looks like runtime_library_dirs is a magic word | 04:36 |
| kota_ | https://docs.python.org/2.7/distutils/apiref.html#distutils.core.Extension | 04:38 |
| kota_ | nice!!!!!!!!! | 04:38 |
| kota_ | runtime_library_dirslist of directories to search for C/C++ libraries at run time (for shared extensions, this is when the extension is loaded) | 04:39 |
| *** psachin has joined #openstack-swift | 04:49 | |
| *** links has joined #openstack-swift | 04:57 | |
| *** tqtran has joined #openstack-swift | 04:59 | |
| *** tqtran has quit IRC | 05:00 | |
| *** ppai has joined #openstack-swift | 05:05 | |
| *** asettle has joined #openstack-swift | 05:06 | |
| *** asettle has quit IRC | 05:11 | |
| *** robcresswell has quit IRC | 05:14 | |
| *** onovy has quit IRC | 05:14 | |
| *** AndyWojo has quit IRC | 05:14 | |
| *** mgagne has quit IRC | 05:15 | |
| *** nottrobin has quit IRC | 05:15 | |
| *** DuncanT_ has joined #openstack-swift | 05:15 | |
| *** samueldmq_ has joined #openstack-swift | 05:15 | |
| *** wasmum has quit IRC | 05:15 | |
| *** CrackerJackMack has quit IRC | 05:15 | |
| *** DuncanT has quit IRC | 05:15 | |
| *** DuncanT_ is now known as DuncanT | 05:16 | |
| *** zacksh has quit IRC | 05:16 | |
| *** jamielennox has quit IRC | 05:16 | |
| *** samueldmq has quit IRC | 05:16 | |
| *** samueldmq_ is now known as samueldmq | 05:17 | |
| *** robcresswell has joined #openstack-swift | 05:17 | |
| *** zacksh has joined #openstack-swift | 05:18 | |
| *** AndyWojo has joined #openstack-swift | 05:19 | |
| *** onovy has joined #openstack-swift | 05:20 | |
| *** wasmum has joined #openstack-swift | 05:20 | |
| *** CrackerJackMack has joined #openstack-swift | 05:20 | |
| *** dmorita has joined #openstack-swift | 05:22 | |
| *** nottrobin has joined #openstack-swift | 05:23 | |
| *** dmorita has quit IRC | 05:26 | |
| *** jamielennox has joined #openstack-swift | 05:31 | |
| *** Jeffrey4l has quit IRC | 05:35 | |
| *** Jeffrey4l has joined #openstack-swift | 05:35 | |
| *** dmorita has joined #openstack-swift | 06:24 | |
| *** SkyRocknRoll has joined #openstack-swift | 06:25 | |
| *** dmorita has quit IRC | 06:28 | |
| *** SkyRocknRoll has quit IRC | 06:30 | |
| *** acorwin_ has quit IRC | 06:32 | |
| *** acorwin has joined #openstack-swift | 06:34 | |
| *** ChubYann has quit IRC | 06:42 | |
| *** sams-gleb has joined #openstack-swift | 06:42 | |
| *** SkyRocknRoll has joined #openstack-swift | 06:43 | |
| timburke | kota_: yeah, but you still need library_dirs to build it in the first place. you need the combo | 06:44 |
| timburke | i ought to look at using sys.prefix, though | 06:44 |
| *** silor has joined #openstack-swift | 06:51 | |
| *** mingyu has joined #openstack-swift | 06:54 | |
| *** silor1 has joined #openstack-swift | 06:56 | |
| *** silor has quit IRC | 06:57 | |
| *** silor1 is now known as silor | 06:57 | |
| *** sams-gleb has quit IRC | 06:59 | |
| *** sams-gleb has joined #openstack-swift | 06:59 | |
| *** sams-gleb has quit IRC | 07:03 | |
| *** winggundamth_ has quit IRC | 07:05 | |
| openstackgerrit | Tim Burke proposed openstack/pyeclib: Add tox environment to test against liberasurecode master https://review.openstack.org/407829 | 07:07 |
| *** asettle has joined #openstack-swift | 07:08 | |
| *** jamielennox is now known as jamielennox|away | 07:11 | |
| kota_ | timburke: perhaps, we could build isa-l from source too with same way for testing at gate? | 07:12 |
| *** asettle has quit IRC | 07:12 | |
| timburke | kota_: yeah, i'd expect we can. hadn't looked into it yet | 07:12 |
| *** sileht has quit IRC | 07:13 | |
| kota_ | timburke: amazing work, absolutely worthful | 07:13 |
| timburke | i wonder if just cloning/installing isa-l first, and using the same --prefix for both might be enough? | 07:14 |
| kota_ | timburke: I think so because you have worked on sys.prefix | 07:14 |
| kota_ | timburke: isa-l installation is really similar way with libersaurecode, autogen, configure, make, make install | 07:15 |
| timburke | (part of it was i realized that we're going to be fighting these same battles repeatedly, so i ought to start looking for some solutions now) | 07:15 |
| *** manous has quit IRC | 07:20 | |
| *** klrmn has quit IRC | 07:21 | |
| *** hseipp has joined #openstack-swift | 07:31 | |
| *** dmorita has joined #openstack-swift | 07:34 | |
| *** rcernin has joined #openstack-swift | 07:34 | |
| *** manous has joined #openstack-swift | 07:37 | |
| *** dmorita has quit IRC | 07:38 | |
| *** pcaruana has joined #openstack-swift | 07:41 | |
| *** mvk has quit IRC | 07:51 | |
| *** Raymii has quit IRC | 07:53 | |
| *** Raymii has joined #openstack-swift | 07:59 | |
| *** zackmdavis has quit IRC | 08:00 | |
| *** zackmdavis has joined #openstack-swift | 08:00 | |
| *** tdasilva has quit IRC | 08:08 | |
| *** asettle has joined #openstack-swift | 08:08 | |
| *** asettle has quit IRC | 08:13 | |
| *** manous has quit IRC | 08:16 | |
| *** Jeffrey4l has quit IRC | 08:17 | |
| *** Jeffrey4l has joined #openstack-swift | 08:17 | |
| *** zul has quit IRC | 08:18 | |
| *** mvk has joined #openstack-swift | 08:26 | |
| *** sileht has joined #openstack-swift | 08:26 | |
| *** amoralej|off is now known as amoralej | 08:29 | |
| *** mingyu has quit IRC | 08:31 | |
| *** zul has joined #openstack-swift | 08:32 | |
| *** tdasilva has joined #openstack-swift | 08:43 | |
| *** dmorita has joined #openstack-swift | 08:46 | |
| *** sorrison_laptop has joined #openstack-swift | 08:51 | |
| *** dmorita has quit IRC | 08:51 | |
| *** daemontool_ has quit IRC | 08:59 | |
| *** sorrison_laptop is now known as sorrison_ | 09:00 | |
| *** Worry has joined #openstack-swift | 09:02 | |
| *** asettle has joined #openstack-swift | 09:09 | |
| *** asettle has quit IRC | 09:14 | |
| *** cbartz has joined #openstack-swift | 09:20 | |
| *** oshritf has joined #openstack-swift | 09:20 | |
| *** jordanP has joined #openstack-swift | 09:22 | |
| *** oshritf has quit IRC | 09:23 | |
| *** oshritf has joined #openstack-swift | 09:27 | |
| *** oshritf has quit IRC | 09:37 | |
| *** NeverWorry has joined #openstack-swift | 09:42 | |
| *** Worry has quit IRC | 09:44 | |
| *** Worry has joined #openstack-swift | 09:44 | |
| *** Worry has quit IRC | 09:45 | |
| *** Worry has joined #openstack-swift | 09:45 | |
| *** NeverWorry has quit IRC | 09:46 | |
| *** asettle has joined #openstack-swift | 09:52 | |
| *** acoles_ is now known as acoles | 09:53 | |
| *** mvk has quit IRC | 09:54 | |
| *** Worry has quit IRC | 09:54 | |
| *** Worry has joined #openstack-swift | 09:55 | |
| *** mvk has joined #openstack-swift | 09:55 | |
| *** tqtran has joined #openstack-swift | 10:01 | |
| *** Worry has quit IRC | 10:02 | |
| *** tqtran has quit IRC | 10:05 | |
| *** abalfour has quit IRC | 10:09 | |
| *** abalfour has joined #openstack-swift | 10:09 | |
| *** sorrison_ has quit IRC | 10:09 | |
| *** DuncanT has quit IRC | 10:33 | |
| *** DuncanT has joined #openstack-swift | 10:33 | |
| *** david-lyle_ has joined #openstack-swift | 10:35 | |
| *** david-lyle has quit IRC | 10:37 | |
| *** mgagne has joined #openstack-swift | 10:47 | |
| *** mgagne is now known as Guest2615 | 10:47 | |
| *** npf has joined #openstack-swift | 10:49 | |
| *** delewis has quit IRC | 10:56 | |
| *** delewis has joined #openstack-swift | 10:58 | |
| *** EmilienM has left #openstack-swift | 11:20 | |
| *** SkyRocknRoll has quit IRC | 11:29 | |
| *** openstackgerrit has quit IRC | 11:32 | |
| *** cdelatte has joined #openstack-swift | 11:34 | |
| *** ppai has quit IRC | 11:37 | |
| *** arch-nemesis has joined #openstack-swift | 11:42 | |
| *** SkyRocknRoll has joined #openstack-swift | 11:44 | |
| *** arch-nemesis has quit IRC | 11:49 | |
| *** ppai has joined #openstack-swift | 11:53 | |
| *** catintheroof has joined #openstack-swift | 11:55 | |
| *** arch-nemesis has joined #openstack-swift | 12:02 | |
| *** tqtran has joined #openstack-swift | 12:02 | |
| *** dmorita has joined #openstack-swift | 12:03 | |
| *** tqtran has quit IRC | 12:07 | |
| *** dmorita has quit IRC | 12:07 | |
| *** vint_bra has joined #openstack-swift | 12:24 | |
| *** catintheroof has quit IRC | 12:32 | |
| *** catintheroof has joined #openstack-swift | 12:32 | |
| *** catintheroof has quit IRC | 12:37 | |
| *** asettle has quit IRC | 12:37 | |
| *** asettle has joined #openstack-swift | 12:38 | |
| *** catintheroof has joined #openstack-swift | 12:54 | |
| *** amoralej is now known as amoralej|lunch | 13:08 | |
| *** catinthe_ has joined #openstack-swift | 13:14 | |
| *** catintheroof has quit IRC | 13:16 | |
| *** tuan_luong has joined #openstack-swift | 13:36 | |
| *** rcernin has quit IRC | 13:49 | |
| *** SkyRocknRoll has quit IRC | 13:50 | |
| *** rcernin has joined #openstack-swift | 13:52 | |
| *** openstackgerrit has joined #openstack-swift | 13:52 | |
| openstackgerrit | Merged openstack/pyeclib: ISA-L Cauchy support https://review.openstack.org/393276 | 13:52 |
| *** Guest2615 is now known as mgagne | 13:54 | |
| *** mgagne has quit IRC | 13:54 | |
| *** mgagne has joined #openstack-swift | 13:54 | |
| *** klamath has joined #openstack-swift | 13:54 | |
| *** bkopilov has quit IRC | 13:59 | |
| *** links has quit IRC | 13:59 | |
| *** bkopilov has joined #openstack-swift | 13:59 | |
| *** catintheroof has joined #openstack-swift | 14:02 | |
| *** manous has joined #openstack-swift | 14:02 | |
| *** tqtran has joined #openstack-swift | 14:04 | |
| *** catinthe_ has quit IRC | 14:04 | |
| *** amoralej|lunch is now known as amoralej | 14:07 | |
| *** tqtran has quit IRC | 14:09 | |
| *** hseipp has quit IRC | 14:12 | |
| *** SkyRocknRoll has joined #openstack-swift | 14:14 | |
| *** SkyRocknRoll has quit IRC | 14:29 | |
| *** Jeffrey4l has quit IRC | 14:30 | |
| *** Jeffrey4l has joined #openstack-swift | 14:31 | |
| *** tuan_luong has quit IRC | 14:38 | |
| *** SkyRocknRoll has joined #openstack-swift | 14:42 | |
| *** StraubTW has joined #openstack-swift | 14:44 | |
| *** tongli has joined #openstack-swift | 14:53 | |
| *** links has joined #openstack-swift | 15:03 | |
| *** chsc has joined #openstack-swift | 15:08 | |
| *** sams-gleb has joined #openstack-swift | 15:12 | |
| *** ppai has quit IRC | 15:23 | |
| *** daemontool has joined #openstack-swift | 15:23 | |
| *** catintheroof has quit IRC | 15:26 | |
| *** catintheroof has joined #openstack-swift | 15:26 | |
| *** Jeffrey4l has quit IRC | 15:27 | |
| *** Jeffrey4l has joined #openstack-swift | 15:27 | |
| *** hseipp has joined #openstack-swift | 15:28 | |
| *** catintheroof has quit IRC | 15:31 | |
| *** dmorita has joined #openstack-swift | 15:38 | |
| *** dmorita has quit IRC | 15:43 | |
| *** isotope has joined #openstack-swift | 15:45 | |
| *** klrmn has joined #openstack-swift | 15:49 | |
| *** hogepodge has joined #openstack-swift | 15:50 | |
| *** manous has quit IRC | 15:52 | |
| *** isotope has quit IRC | 15:55 | |
| *** catintheroof has joined #openstack-swift | 15:57 | |
| *** chlong has joined #openstack-swift | 15:58 | |
| *** tongli has quit IRC | 16:03 | |
| *** links has quit IRC | 16:13 | |
| *** SkyRocknRoll has quit IRC | 16:14 | |
| *** hseipp has quit IRC | 16:14 | |
| openstackgerrit | Sachin Patil proposed openstack/swift: Fix swift-get-nodes arg parsing for missing ring https://review.openstack.org/406012 | 16:17 |
| *** pcaruana has quit IRC | 16:18 | |
| *** rcernin has quit IRC | 16:19 | |
| *** psachin has quit IRC | 16:26 | |
| *** mvk has quit IRC | 16:31 | |
| *** _JZ_ has joined #openstack-swift | 16:33 | |
| *** xionchen_ has joined #openstack-swift | 16:34 | |
| *** MooingLemur has quit IRC | 16:47 | |
| *** chlong has quit IRC | 16:56 | |
| *** asettle has quit IRC | 17:07 | |
| *** rcernin has joined #openstack-swift | 17:11 | |
| *** klrmn has quit IRC | 17:16 | |
| *** dmorita has joined #openstack-swift | 17:21 | |
| *** mvk has joined #openstack-swift | 17:24 | |
| *** MooingLemur has joined #openstack-swift | 17:29 | |
| *** chlong has joined #openstack-swift | 17:30 | |
| *** dmorita has quit IRC | 17:32 | |
| *** dmorita has joined #openstack-swift | 17:32 | |
| *** calebb has quit IRC | 17:34 | |
| *** daemontool has quit IRC | 17:35 | |
| *** dmorita has quit IRC | 17:37 | |
| clayg | zomg zomg zomg did isa-l-rs-cauchy land! ?!!! wooooo! | 17:45 |
| clayg | tdasilva: notmyname: kota: RE-LEASE RE-LEASE RE-LEASE | 17:46 |
| clayg | I think last time we did a liberasure/pyeclib release I lied and said something stupid like "next time you guys should teach me how to do it" - i'm such a liar | 17:47 |
| tdasilva | lol | 17:48 |
| tdasilva | clayg: we want to release both right? | 17:48 |
| tdasilva | it will be coincidentally 1.3.2 of each | 17:48 |
| clayg | zomg zomg zomg - i want them so much - i could hug you! | 17:49 |
| clayg | last time with pyeclib i remeber it took two tries for some reason? so... i don't if it would be useful to look at the history of those releases and double check all the p's and q's - i have it on good authority that software is hard | 17:50 |
| patchbot | hard | 17:50 |
| timburke | i feel like https://review.openstack.org/#/c/405926/ might be nice to get in | 17:50 |
| patchbot | patch 405926 - liberasurecode - Add get_version functionality to liberasurecode lib | 17:50 |
| clayg | timburke: that's the ticket! | 17:50 |
| clayg | timburke: how do i sanity check it - do you have a comment in your review like "when I fobrinate the wizbang - boom" | 17:51 |
| timburke | like, what's the point of using a dynamically loaded library if you can't actually know which version you're loading? | 17:51 |
| clayg | i feel like it would maybe be something like build something and then look at some attirbute on a module somewhere before and after? | 17:52 |
| zaitcev | I did not know "rescure" was a word | 17:52 |
| timburke | i could whip something up... hold on... | 17:52 |
| zaitcev | wait, it's a typo | 17:52 |
| clayg | oh... hrmm... this is just the liberasurecode side - so it'd be a little 'test.c' thingy - is there a pyeclib side that goes with it!? | 17:52 |
| clayg | cause I will totally review end-to-end | 17:53 |
| timburke | zaitcev: or is it ... "what supertoast would do"? what?? http://www.urbandictionary.com/define.php?term=rescure | 17:53 |
| tdasilva | clayg, timburke: while you look at patch 405926 i will start loading release procedures into my brain again and get ready to release when you guys say go | 17:54 |
| patchbot | https://review.openstack.org/#/c/405926/ - liberasurecode - Add get_version functionality to liberasurecode lib | 17:54 |
| clayg | tdasilva: i don't think it matters that much if the released pyeclib isn't using it | 17:54 |
| clayg | maybe we should *just* have liberasurecode - which is obstentially a c project that provides .so's but that *also* happens to have just a *tad* bith of python in it to provide python bindings and may some day have bindings for another compiled language - can't think off the top of my head which language that might be... | 17:55 |
| clayg | *ostensibly - i can't even *say* that word - much less spell it | 17:56 |
| timburke | clayg: https://review.openstack.org/#/c/405929/ iirc | 17:56 |
| patchbot | patch 405929 - pyeclib - WIP: Change the version reference | 17:56 |
| clayg | timburke: that's the ticket! | 17:56 |
| * clayg votes timburke for google in the next election for "keeper of all worlds knowledge" | 17:56 | |
| *** dmorita has joined #openstack-swift | 17:57 | |
| clayg | timburke: except patch 405929 can't merge because it tests a new function on liberasurecode master | 17:58 |
| patchbot | https://review.openstack.org/#/c/405929/ - pyeclib - WIP: Change the version reference | 17:58 |
| clayg | the gate unittest failure is legit :\ | 17:58 |
| clayg | timburke: so we're back to your make tox install liberasurecode from source .. and also probably update the gate .. and also probably break everyone's dev environments if we try to *use* the new pyeclib function from swift | 17:59 |
| clayg | timburke: I don't think we *should* try to get this in right now? it doesn't fix any critical production issues for me and it seems... hard? | 18:00 |
| *** calebb has joined #openstack-swift | 18:00 | |
| clayg | timburke: also instead of compiling from source I'm going for drop pyeclib and move python bindings for liberasurecode into the liberasurecode project and change swift to depend on python-liberasurecode instead of pyeclib | 18:00 |
| timburke | yeah, and the more i think about it, the more i realize that we probably need to add a #define GET_VERSION_AVAILABLE or something on the liberasure side so we actually know whether that functions is defined. maybe better to hold off | 18:00 |
| tdasilva | timburke, clayg: at some point i was thinking about creating a new gate job for pyeclib where we would build liberasurecode from source, is that what you did with your tox change? | 18:00 |
| clayg | ... until someone tells me that's a bad idea | 18:01 |
| clayg | and i'm not worrying about it today | 18:01 |
| timburke | tdasilva: yup. playing with getting isa-l installed, too | 18:01 |
| tdasilva | ah ok | 18:01 |
| tdasilva | cool! | 18:01 |
| clayg | tdasilva: he created the tox job - not the gate job - but that should make the gate job easier (!?) | 18:01 |
| *** david-lyle_ is now known as david-lyle | 18:02 | |
| clayg | timburke: tdasilva: no whey - get on board with this fork pyeclib into a subdirectory of liberasure code train - it's going to be the bomb! | 18:02 |
| tdasilva | pyeclib is just a python binding of liberasurecode, right? | 18:03 |
| tdasilva | why not? | 18:03 |
| clayg | yeah it makes sense to me - and also when we have golang bindings for liberasurecode - as if we're going to be able to keep *three* repos stright? | 18:03 |
| clayg | add a thing to liberasurecode c - then expose it in python and golang bindings - all three unittests pass - merge it, build it, test it, ship it, $$$ | 18:04 |
| tdasilva | clayg: if we ever wanted to make use of another ec library like https://memoscale.com/ | 18:04 |
| tdasilva | that would be plugged in to liberasurecode or pyeclib? | 18:04 |
| tdasilva | actually NTT also has a ec library right? | 18:05 |
| timburke | shss | 18:05 |
| tdasilva | yeah, that | 18:05 |
| clayg | for us - and the way we deal with ec - i'm guessing they'd plug into liberasurecode - and then we could expose it in the python or golang bindings at the same time | 18:06 |
| clayg | all in one repo | 18:06 |
| tdasilva | clayg: sounds like a good idea to me | 18:06 |
| clayg | we couldn't consume it directly from swift - even if they have python and golang bindings - unless we made a shim to match the liberasurecode/pyeclib interface | 18:06 |
| clayg | but yeah - just like isa-l or ssh - it'd be easy to have liberasurecode be the shim that wraps up the other ec library to a common interface and then exposed in all the language bindings that liberasurecode supports | 18:07 |
| clayg | I think someone thought if liberasurecode grew lots of bindings for different languages - it would be easier if they were seperate repos - but I think ... i'm pretty sure ... based on experience and hindsight - this is not easy | 18:08 |
| clayg | it's possible the other way may *also* be not easy | 18:08 |
| clayg | grass is greener kind of problem | 18:08 |
| *** cbartz has quit IRC | 18:09 | |
| clayg | tdasilva: but for now - let's cut the release - and maybe we can discuss more with kota and think about making a move in Jan? | 18:10 |
| *** cnf has quit IRC | 18:10 | |
| clayg | tdasilva: timburke and me think maybe the version thing might need a few more revs | 18:10 |
| *** calebb has quit IRC | 18:11 | |
| *** klrmn has joined #openstack-swift | 18:11 | |
| *** cnf has joined #openstack-swift | 18:11 | |
| tdasilva | clayg: ok, sounds good | 18:12 |
| *** isotope has joined #openstack-swift | 18:18 | |
| *** asettle has joined #openstack-swift | 18:21 | |
| *** amoralej is now known as amoralej|off | 18:25 | |
| *** rickflare has quit IRC | 18:26 | |
| *** rickflare has joined #openstack-swift | 18:31 | |
| tdasilva | clayg, timburke: libec adding support for isa-l cauchy, makes it a 1.3.*2* release or a 1.*4*.0 release? | 18:32 |
| tdasilva | thoughts? | 18:32 |
| timburke | i'm partial towards 1.4.0 | 18:34 |
| *** chlong has quit IRC | 18:38 | |
| openstackgerrit | Thiago da Silva proposed openstack/liberasurecode: 1.4.0 release https://review.openstack.org/408257 | 18:40 |
| tdasilva | timburke: me too :) | 18:40 |
| tdasilva | although i changed my mind after creating branch name | 18:40 |
| *** mcarden has quit IRC | 18:40 | |
| tdasilva | and forgot to rename it | 18:40 |
| *** mcarden has joined #openstack-swift | 18:41 | |
| *** clu_ has joined #openstack-swift | 18:45 | |
| openstackgerrit | Tim Burke proposed openstack/pyeclib: Add tox environment to test against liberasurecode master https://review.openstack.org/407829 | 18:48 |
| openstackgerrit | Tim Burke proposed openstack/pyeclib: Install isa-l from source when testing liberasure-master https://review.openstack.org/408259 | 18:48 |
| openstackgerrit | Thiago da Silva proposed openstack/pyeclib: renaming README.md to README.rst https://review.openstack.org/408260 | 18:50 |
| tdasilva | timburke: i'd like to get that last patch 408260 in before release | 18:51 |
| patchbot | https://review.openstack.org/#/c/408260/ - pyeclib - renaming README.md to README.rst | 18:51 |
| openstackgerrit | Tim Burke proposed openstack/pyeclib: Add tests for the availability of individual backends https://review.openstack.org/408261 | 18:51 |
| tdasilva | pypi doesn't support markdown, so nothing gets displayed here https://pypi.python.org/pypi/PyECLib/1.3.1 | 18:51 |
| timburke | looking now. was it really already rst? no conversion necessary? | 18:52 |
| *** jordanP has quit IRC | 18:52 | |
| tdasilva | timburke: i only quickly checked with an atom preview package i have and it looked ok | 18:55 |
| *** krypto has joined #openstack-swift | 18:55 | |
| *** chlong has joined #openstack-swift | 18:57 | |
| tdasilva | timburke: http://rst.ninjs.org/?n=33b8954acf3b589b07ac528062b6b490&theme=basic | 18:57 |
| krypto | hello all i have one vm as swift server with 1 Tb disk in the environment and planning to add 5 new physical servers each with 6x1 TBdisks and then slowly remove the vm from the environment.Is this possible ie a cluster with un even number of disks and then removing node from that cluste | 18:59 |
| *** dmorita has quit IRC | 19:04 | |
| *** dmorita has joined #openstack-swift | 19:05 | |
| tdasilva | krypto: initially that plan looks ok IMO, but in your one vm with one disk do you have only a single replica ring? | 19:14 |
| krypto | tdasilva if i may i ask ,how can i check the rings .. honetsly its deployed by somebody else | 19:18 |
| tdasilva | krypto: do you have access to your vm? like ssh access? | 19:19 |
| timburke | clayg: tdasilva: kota_: oh man, metaclasses are great! look at that: http://logs.openstack.org/61/408261/1/check/gate-pyeclib-python27-ubuntu-xenial/ac635f4/console.html#_2016-12-07_18_53_37_979203 beautiful! somebody should go merge https://review.openstack.org/#/c/408261/ like right now ;-) | 19:19 |
| patchbot | patch 408261 - pyeclib - Add tests for the availability of individual backends | 19:19 |
| krypto | tdasilva yes i do | 19:19 |
| *** silor has quit IRC | 19:20 | |
| *** asettle has quit IRC | 19:21 | |
| *** tqtran has joined #openstack-swift | 19:26 | |
| tdasilva | krypto: the rings are in /etc/swift, you can use the swift-ring-builder tool to check the rings | 19:28 |
| tdasilva | krypto: it's probably a good idea to checkout the docs here: http://docs.openstack.org/developer/swift/overview_ring.html http://docs.openstack.org/developer/swift/admin_guide.html | 19:28 |
| tdasilva | there's a ton of detail about the rings | 19:29 |
| krypto | thanks i will check those | 19:31 |
| *** AbyssOne has quit IRC | 19:33 | |
| tdasilva | timburke: thanks for the review on the README, what i'm going to do is to abandon that change and make all the changes in one "release" commit | 19:33 |
| tdasilva | that will contain the changelog and version updates | 19:33 |
| tdasilva | that ok with you? | 19:33 |
| timburke | tdasilva: sure | 19:34 |
| tdasilva | timburke: ok, thanks, sorry for the noise | 19:34 |
| *** acoles is now known as acoles_ | 19:35 | |
| *** AbyssOne has joined #openstack-swift | 19:36 | |
| *** chlong has quit IRC | 19:36 | |
| *** npf has quit IRC | 19:37 | |
| *** npf has joined #openstack-swift | 19:37 | |
| tdasilva | timburke: https://review.openstack.org/#/c/408260/1/README.rst@40 do you mean liberasurecode_rs_cauchy or isa_l_rs_cauchy ? | 19:41 |
| patchbot | patch 408260 - pyeclib - renaming README.md to README.rst | 19:41 |
| *** ChubYann has joined #openstack-swift | 19:41 | |
| timburke | tdasilva: bah, i meant isa_l_rs_cauchy -- you're right | 19:42 |
| *** StraubTW has quit IRC | 19:45 | |
| *** StraubTW has joined #openstack-swift | 19:46 | |
| *** npf has quit IRC | 19:52 | |
| *** calebb has joined #openstack-swift | 19:58 | |
| *** chlong has joined #openstack-swift | 20:00 | |
| openstackgerrit | Thiago da Silva proposed openstack/pyeclib: release 1.4.0 https://review.openstack.org/408280 | 20:02 |
| timburke | tdasilva: on footnotes, i mean things that link to the bit at the bottom, like in http://docutils.sourceforge.net/docs/user/rst/quickref.html#footnotes | 20:02 |
| tdasilva | timburke: oh yeah, i understand now | 20:04 |
| *** xionchen_ has quit IRC | 20:04 | |
| tdasilva | timburke: i think that whole README needs a bit of reformatting, it looks quite different from a typical README one finds on github. could we leave that for after this release? | 20:05 |
| timburke | tdasilva: sure. i think as long as we add the new feature, i'm happy | 20:05 |
| tdasilva | :) | 20:06 |
| tdasilva | kota_, clayg, timburke, notmyname : patch 408257 and patch 408280 are ready for review, we should be able to tag the release after these two merge | 20:09 |
| patchbot | https://review.openstack.org/#/c/408257/ - liberasurecode - 1.4.0 release | 20:09 |
| patchbot | https://review.openstack.org/#/c/408280/ - pyeclib - release 1.4.0 | 20:09 |
| *** dmorita has quit IRC | 20:14 | |
| *** dmorita has joined #openstack-swift | 20:16 | |
| openstackgerrit | Tim Burke proposed openstack/pyeclib: Install jerasure from source when testing liberasure-master https://review.openstack.org/408286 | 20:16 |
| timburke | that's the stuff. https://review.openstack.org/#/c/408286/ (and the two parent commits in its chain) + https://review.openstack.org/#/c/408261/ = win | 20:18 |
| patchbot | patch 408286 - pyeclib - Install jerasure from source when testing liberasu... | 20:18 |
| patchbot | patch 408261 - pyeclib - Add tests for the availability of individual backends | 20:18 |
| timburke | kota_: if only i knew how to install shss :-( | 20:18 |
| *** dmorita has quit IRC | 20:18 | |
| *** dmorita has joined #openstack-swift | 20:20 | |
| *** isotope has quit IRC | 20:23 | |
| *** asettle has joined #openstack-swift | 20:25 | |
| *** isotope has joined #openstack-swift | 20:26 | |
| timburke | tdasilva: do we want to fix one more memory leak ahead of the pyeclib release? https://review.openstack.org/#/c/407302/ | 20:29 |
| patchbot | patch 407302 - pyeclib - Fix checksum memory leak | 20:29 |
| *** asettle has quit IRC | 20:30 | |
| *** arch-nemesis has quit IRC | 20:30 | |
| *** diogogmt has joined #openstack-swift | 20:43 | |
| *** m_kazuhiro has joined #openstack-swift | 20:47 | |
| *** dmorita has quit IRC | 20:52 | |
| *** dmorita has joined #openstack-swift | 20:52 | |
| *** isotope has quit IRC | 20:52 | |
| *** tqtran is now known as tqtran-afk | 20:59 | |
| tdasilva | oh, no meeting today, right? | 21:03 |
| mathiasb | tdasilva: that's what I just remembered as well | 21:03 |
| *** hseipp has joined #openstack-swift | 21:04 | |
| dmorita | Actually, I checked last week's log, then notmyname said no meeting today. | 21:08 |
| *** asettle has joined #openstack-swift | 21:09 | |
| *** asettle has quit IRC | 21:09 | |
| *** asettle has joined #openstack-swift | 21:10 | |
| *** vint_bra has quit IRC | 21:10 | |
| *** asettle has quit IRC | 21:12 | |
| *** asettle has joined #openstack-swift | 21:12 | |
| *** manous has joined #openstack-swift | 21:13 | |
| tdasilva | timburke: it seems like a good idea to hold up for that memory leak fix. what does clayg think? | 21:18 |
| timburke | tdasilva: we'll find out :-) i think he's on his way in currently | 21:18 |
| timburke | on the plus side, it's a fairly small memory leak; we were having trouble detecting it with the 1000x iterations | 21:19 |
| *** m_kazuhiro has quit IRC | 21:20 | |
| *** a1|away has quit IRC | 21:21 | |
| MooingLemur | Has there been any movement on the ability to change storage policies of existing containers (background migration)? | 21:21 |
| tdasilva | MooingLemur: i think that's something that dmorita was working on....but i'm not sure about the current status | 21:23 |
| dmorita | MooingLemur: yeah, I am working on it. I am now rewriting patch to work this feature with fastPost. Concept is done. But some unit tests do not pass. Therefore, I am now checking some regressions. | 21:25 |
| dmorita | MooingLemur: I think I can update my working patch by next week. | 21:26 |
| MooingLemur | That's great news :) Thanks for all your efforts. I have some... mistakes... to correct :) I have a ring with too high a part power I'd like to migrate away from. | 21:27 |
| *** sams-gleb has quit IRC | 21:32 | |
| *** a1|away has joined #openstack-swift | 21:32 | |
| *** sams-gleb has joined #openstack-swift | 21:33 | |
| *** Jeffrey4l has quit IRC | 21:35 | |
| *** Jeffrey4l has joined #openstack-swift | 21:35 | |
| *** catintheroof has quit IRC | 21:37 | |
| *** sams-gleb has quit IRC | 21:37 | |
| clayg | timburke: where is it - let me at it1 | 21:38 |
| *** catintheroof has joined #openstack-swift | 21:38 | |
| timburke | clayg: https://review.openstack.org/#/c/407302/ | 21:38 |
| patchbot | patch 407302 - pyeclib - Fix checksum memory leak | 21:38 |
| tdasilva | timburke: fwiw i had to increate to 4M to see the failure | 21:39 |
| clayg | lol @ loop 400K | 21:39 |
| patchbot | https://review.openstack.org/#/c/400/ | 21:39 |
| clayg | wtf patchbot - you're drunk | 21:39 |
| timburke | i think it's the p. like p 407302 | 21:39 |
| patchbot | https://review.openstack.org/#/c/407302/ - pyeclib - Fix checksum memory leak | 21:39 |
| timburke | yeah... | 21:39 |
| tdasilva | timburke: well...4M was the first line i tried after the failure didn't happen at 400k | 21:40 |
| tdasilva | s/first line/first number | 21:40 |
| timburke | tdasilva: hrm. we can bump it higher... but if we start getting test times like swift's, maybe we're better off with the subtle memory leaks... | 21:41 |
| clayg | rofl | 21:42 |
| tdasilva | hehehe | 21:42 |
| *** catintheroof has quit IRC | 21:42 | |
| tdasilva | actually, error that i see is with something else, but I did notice that looping 400K times without the actual fix did not show any errors for me | 21:43 |
| clayg | tdasilva: you have to `sudo pip install -e .` when you change the c code to make with the rebuild or something | 21:45 |
| clayg | it def pops for me @ 400K | 21:45 |
| clayg | ... but not at 1K - today 2-3K seems to the ticket | 21:46 |
| tdasilva | clayg: i was running `sudo python setup.py develop` ? is that the same? | 21:46 |
| timburke | clayg: but you and i are probably using rather similar vsaio environments. i wonder what kota_ typically uses... | 21:47 |
| clayg | tdasilva: they *are* the same - unless your setup.py uses pbr - then your command is a good way to traceback while newish pip will mostly work | 21:47 |
| clayg | tdasilva: so you should replace `pip install -e .` in your brain everytime you normally type `python setup.py develop` | 21:48 |
| *** abalfour has left #openstack-swift | 21:48 | |
| clayg | `pip install .` is good for `python setup.py install` | 21:48 |
| clayg | if you need `pyton setup.py sdist` you're screwed - because no one cares about distro packagers anymore | 21:48 |
| *** jamielennox|away is now known as jamielennox | 21:49 | |
| clayg | ok, that's going in | 21:49 |
| tdasilva | clayg: ok, thanks for the heads up | 21:50 |
| timburke | clayg: did you try it for each of test/, test/test_pyeclib_api.py, test/test_pyeclib_api.py:TestPyECLibDriver, and test/test_pyeclib_api.py:TestPyECLibDriver.test_get_metadata_memory_usage ? the test ought to fail regardless of scope; i saw some numbers that would fail for all of tests/ but not test/test_pyeclib_api.py:TestPyECLibDriver, say | 21:50 |
| openstackgerrit | Merged openstack/pyeclib: Fix checksum memory leak https://review.openstack.org/407302 | 21:51 |
| clayg | timburke: i was mostly focused on just test/test_pyeclib_api.py:TestPyECLibDriver.test_get_metadata_memory_usage to ensure it was stable in isolation - i'm happy enough that it's not going to be horribly problematic - "definately better; probably good enough" | 21:53 |
| *** hseipp has quit IRC | 21:57 | |
| mattoliverau | Morning | 21:57 |
| tdasilva | mattoliverau: o/ | 21:59 |
| *** asettle has quit IRC | 22:10 | |
| *** cnf has quit IRC | 22:10 | |
| timburke | clayg: the 40M iterations thing -- was it a MemoryError on loop_range = range(40000000)? | 22:11 |
| clayg | timburke: ain't nobody got time for that! i didn't test it | 22:12 |
| *** cnf has joined #openstack-swift | 22:12 | |
| *** tqtran-afk is now known as tqtran | 22:13 | |
| timburke | oh, it was just speculation; got it | 22:13 |
| *** clu_ has quit IRC | 22:13 | |
| clayg | ah, yeah - sorry - i see now that could be read either way - although if it failed for me a 40M i probably (heh, there it is again) would have said "this failed for me when I tried it at 40M, but it's NBD IMHO" | 22:15 |
| *** clu_ has joined #openstack-swift | 22:15 | |
| clayg | ok timburke are you done fixing all the bugs now? can tdasilva make a tarball for me!? :D | 22:16 |
| * clayg imagines timburke saying "well *acctually* there is this one patch that fixes this other critical bug that's been open for months..." | 22:16 | |
| timburke | clayg: i think? the other stuff i've got out there is mainly test infrastructure | 22:16 |
| pdardeau | timburke: at first i read that ^ as 'manly' test infrastructure. now i gotta know - is it manly? ;-) | 22:18 |
| timburke | clayg: 400M seems to check out fine, so i'm calling it good? | 22:19 |
| clayg | timburke: definately better - probably good enough | 22:21 |
| timburke | pdardeau: well, the chain ending in https://review.openstack.org/#/c/408286/ is a little ugly & brutish, so maybe? i liked writing https://review.openstack.org/#/c/408261/ more | 22:21 |
| patchbot | patch 408286 - pyeclib - Install jerasure from source when testing liberasu... | 22:21 |
| patchbot | patch 408261 - pyeclib - Add tests for the availability of individual backends | 22:21 |
| clayg | install jerasure from source! like pirates!? | 22:22 |
| timburke | clayg: idk? there seems to be a public git server willing to send me files? *i'm* not doing any distributing, at any rate | 22:23 |
| clayg | hehhehe | 22:25 |
| timburke | clayg: http://jerasure.org/jerasure/jerasure/blob/master/License.txt looks like 3-clause BSD | 22:25 |
| timburke | as does http://jerasure.org/jerasure/gf-complete/blob/master/License.txt | 22:29 |
| timburke | ooooh, *patent* issues. yeah, i got nothing there. seems like the isa-l stuff might be of more interest anyway? w/e, it's the last in the chain, we can drop it | 22:32 |
| *** dmorita has quit IRC | 22:33 | |
| *** dmorita has joined #openstack-swift | 22:33 | |
| clayg | timburke: umm... IANAL - so I don't really speak on those things - i have no idea about any of it really - i just know that I don't use jerasure and don't pay much attention to it | 22:36 |
| *** thurloat has quit IRC | 22:39 | |
| clayg | timburke: umm... IANAL - so I don't really speak on those things - i have no idea about any of it really - i just know that I don't use jerasure and don't pay much attention to it | 22:45 |
| clayg | gah! | 22:45 |
| *** rcernin has quit IRC | 22:48 | |
| *** thurloat has joined #openstack-swift | 22:50 | |
| *** cnf has quit IRC | 23:01 | |
| *** cnf has joined #openstack-swift | 23:04 | |
| *** StraubTW has quit IRC | 23:06 | |
| *** chlong has quit IRC | 23:06 | |
| *** dmorita has quit IRC | 23:07 | |
| *** dmorita has joined #openstack-swift | 23:08 | |
| *** asettle has joined #openstack-swift | 23:11 | |
| *** asettle has quit IRC | 23:16 | |
| *** jamielennox is now known as jamielennox|away | 23:27 | |
| *** jamielennox|away is now known as jamielennox | 23:28 | |
| clayg | tdasilva: there's probalby some code that has to merge before a tag show up on at https://github.com/openstack/liberasurecode/releases right? | 23:36 |
| *** klamath has quit IRC | 23:53 | |
| timburke | oh, clayg, here's that demo code for get_version -- http://paste.openstack.org/show/591738/ | 23:56 |
| timburke | i installed liberasurecode to some junk user dir like tmp, then checked that it worked correctly with something like `gcc test.c -L tmp/lib -I tmp/include -l erasurecode -l dl -Xlinker -rpath -Xlinker $(pwd)/tmp/lib && ./a.out` | 23:57 |
| timburke | after that, i played around with switching between before/after kota_'s change or changing the numbers in erasurecode_version.h then running `make && make install && ./a.out`, and even blew away all of tmp/lib and running it again | 23:57 |
| timburke | i stand by my +2 | 23:57 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!